Sudden Cardiac Death and Risk Stratification
Sudden cardiac death kills more people each year than any individual cancer. The EP question is whether we can identify who is at risk before the first event.
The mechanism in the vast majority of sudden cardiac death (SCD) is ventricular tachycardia degenerating into ventricular fibrillation.
The sequence is consistent. Diseased ventricular myocardium generates a reentrant VT. Within minutes, the organized circuit fragments into the chaotic, self-sustaining wavelets of VF. Cardiac output drops to zero. Without defibrillation, survival is near zero.
We have spent the preceding chapters asking when to ablate, when to implant a device, when to treat medically, and when to observe. SCD is the reason devices exist. The entire field of risk stratification exists to answer one question: which patients carry enough substrate to generate lethal VT/VF, and can we identify them before the arrhythmia occurs?
The Mechanism of Sudden Death
SCD begins with VT initiation in vulnerable substrate. The substrate varies by disease: dense scar from a prior myocardial infarction, patchy fibrosis in dilated cardiomyopathy, myocardial disarray in hypertrophic cardiomyopathy, or ion channel dysfunction in the channelopathies. In each case, the common pathway is the same. A premature ventricular complex (PVC) encounters heterogeneous tissue, initiates reentry, and VT begins.
Organized VT can be tolerated briefly. The ventricle contracts in a coordinated pattern, producing some cardiac output, even at rates of 180 to 220 bpm. The danger lies in degeneration. As VT continues, ischemia worsens, the wavefront encounters increasingly heterogeneous refractoriness, and the single reentrant circuit fractures into multiple wavelets. This is the transition from VT to VF. Once VF is established, effective contraction ceases. The ventricle quivers instead of pumping.
The window from VT onset to circulatory collapse may be seconds to minutes. For every minute without defibrillation, survival drops roughly 7 to 10%. By 10 minutes, meaningful neurologic recovery is rare. This is why the ICD exists: to deliver therapy within seconds.
Primary vs. Secondary Prevention
The distinction between primary prevention and secondary prevention defines two fundamentally different clinical problems.
Secondary Prevention
The patient already survived an arrest or sustained VT
The substrate has already declared itself. The patient had a cardiac arrest from VT/VF or documented sustained VT with hemodynamic compromise. Recurrence risk is high: without an ICD, approximately 30% of survivors will have another event within two years.
An ICD is indicated in nearly all cases, provided the event was not caused by a completely reversible trigger (such as the first 48 hours of an acute MI or a corrected electrolyte abnormality). The evidence here is straightforward. These patients have proven they have lethal substrate.
Primary Prevention
The patient has never had an event
The patient has substrate features that place them at elevated risk, but has never had a cardiac arrest or sustained VT. The question is whether the risk of a future event justifies the burden and complications of an ICD.
This is where risk stratification matters. We need tools that identify patients whose substrate is dangerous enough to warrant prophylactic defibrillator implantation. The challenge: the individual risk in this population is lower, and many implanted devices will never fire an appropriate shock. We are treating a population to save a subset.
Ejection Fraction: The Dominant Risk Marker
Left ventricular ejection fraction (LVEF) ≤ 35% remains the single most widely used criterion for primary prevention ICD implantation. The evidence comes from two landmark trials.
MADIT-II enrolled patients with ischemic cardiomyopathy and EF ≤ 30%. ICD implantation reduced all-cause mortality by 31% over 20 months. SCD-HeFT extended this to both ischemic and non-ischemic cardiomyopathy patients with EF ≤ 35% and NYHA class II or III heart failure. The ICD arm showed a 23% relative reduction in death over 5 years.
The mechanistic logic is straightforward. Low EF reflects extensive myocardial disease. More disease means more scar, more fibrosis, more regions of slow conduction and heterogeneous refractoriness. This is precisely the substrate that supports reentrant VT. A heart with 25% EF from a prior anterior MI has a large scar with surviving muscle bundles threading through it. Those bundles are the reentrant circuits waiting to activate.
EF is a blunt instrument. It correlates with scar burden at the population level, but it misses critical nuance at the individual level. Most SCDs in absolute numbers occur in patients with EF > 35%. These patients have smaller scars, focal fibrosis, or channelopathies that EF cannot detect. Meanwhile, many patients with EF ≤ 35% receive ICDs that never deliver an appropriate shock. They bear the risks of device complications (infection, lead failure, inappropriate shocks) without receiving the benefit.
EF tells us the ventricle is sick. It does not tell us whether the specific pattern of disease supports lethal reentry.
› Deep Dive: The Debate Over Primary Prevention ICD Criteria
LVEF ≤ 35% became the standard ICD threshold because MADIT-II and SCD-HeFT demonstrated mortality reduction in this population. But EF is a measure of pump function, not arrhythmic substrate. Two patients with EF of 30% can have vastly different scar architectures: one with a thin, well-defined anterior scar and minimal border zone, the other with extensive patchy fibrosis and multiple potential reentry circuits. EF treats them identically.
The numbers reveal the problem. In SCD-HeFT, the ICD arm showed a 7.2% absolute mortality reduction at 5 years, meaning roughly 14 patients needed to be implanted to save one life (NNT ≈ 14). The remaining 13 patients carried the risks of device complications without receiving the arrhythmic benefit. Meanwhile, in the general population, the majority of SCDs in absolute numbers occur in patients whose EF exceeds 35%, placing them outside current implantation guidelines.
Ongoing research aims to refine risk prediction beyond EF. Cardiac MRI scar quantification (total scar mass, border zone volume) identifies substrate features that EF cannot detect. The CMR-Guide trial is testing whether MRI-guided ICD implantation in patients with non-ischemic cardiomyopathy and EF 36-50% can capture high-risk patients currently missed by EF criteria. Other markers under investigation include programmed stimulation results, autonomic function testing, genetic risk scores, and machine-learning models integrating ECG, imaging, and clinical variables. The goal is a risk stratification framework that moves from a single threshold (EF ≤ 35%) to a multi-parameter assessment of each patient's individual substrate.
Beyond EF: Refining Risk
Because EF alone is insufficient, multiple complementary tools exist to characterize the arrhythmic substrate more precisely.
Late gadolinium enhancement (LGE) on cardiac MRI directly visualizes scar tissue. Gadolinium accumulates in fibrotic myocardium and washes out of healthy tissue, producing bright signal in areas of scar. Total scar mass correlates with VT risk. More importantly, the border zone, the region of intermediate signal intensity between dense scar and normal myocardium, represents the heterogeneous tissue where reentrant circuits form. Border zone volume predicts VT inducibility at EP study and SCD risk independently of EF.
The signal-averaged ECG (SAECG) amplifies low-amplitude signals at the terminal portion of the QRS complex. These late potentials represent slow conduction through scarred myocardium, the same substrate that supports reentry. Their presence after MI identifies patients with higher VT risk. Their absence has good negative predictive value.
Non-sustained VT (NSVT) on Holter monitoring identifies patients with irritable substrate. In the setting of reduced EF, NSVT increases SCD risk several-fold. Three or more consecutive ventricular beats at ≥100 bpm that terminate spontaneously before 30 seconds: the substrate can initiate VT, but the conditions for sustained reentry were not met during that episode.
Microvolt T-wave alternans (MTWA) detects beat-to-beat fluctuation in T-wave morphology during exercise. This alternation reflects spatial dispersion of repolarization, the same heterogeneity that allows a premature beat to find tissue in different refractory states and initiate reentry. A positive test indicates elevated risk. A negative test at adequate heart rate has strong negative predictive value.
Programmed ventricular stimulation delivers premature extrastimuli to test whether the substrate supports reentry. If sustained monomorphic VT is inducible, the patient has a circuit that can be activated. In ischemic cardiomyopathy, inducible VT significantly increases SCD risk and strengthens the case for ICD implantation. In non-ischemic cardiomyopathy, the predictive value is lower because the arrhythmia mechanism is more heterogeneous.
Special Populations
Several conditions carry elevated SCD risk through mechanisms distinct from ischemic scar. Each requires its own risk stratification framework.
HCM
Hypertrophic Cardiomyopathy
The substrate in hypertrophic cardiomyopathy (HCM) is myocardial disarray: disorganized myocyte architecture with interspersed fibrosis, typically concentrated in the hypertrophied septum. Sarcomere gene mutations (most commonly in beta-myosin heavy chain and myosin-binding protein C) drive the hypertrophy, but it is the resulting microscopic chaos that matters electrically. Disarray creates heterogeneous conduction at the cellular level, supporting reentry even without macroscopic scar.
Established risk factors for SCD in HCM include:
- Massive hypertrophy (wall thickness ≥30 mm)
- Family history of SCD from HCM
- Unexplained syncope
- Non-sustained VT on ambulatory monitoring
- Abnormal blood pressure response during exercise (failure to augment or a drop in systolic pressure)
- Extensive LGE on cardiac MRI (≥15% of LV mass)
Current guidelines use risk calculators (such as the HCM Risk-SCD model) that integrate multiple variables to estimate 5-year SCD probability.
ARVC
Arrhythmogenic Right Ventricular Cardiomyopathy
In arrhythmogenic right ventricular cardiomyopathy (ARVC), desmosomal protein mutations cause progressive fibrofatty replacement of the right ventricular myocardium. The surviving muscle fibers within this replacement tissue create the substrate for reentrant VT, often with a left bundle branch block morphology (arising from the RV).
Diagnosis follows the revised Task Force criteria, combining structural, histological, ECG, arrhythmic, and genetic features. Risk stratification uses dedicated calculators incorporating sustained VT, syncope, NSVT, PVC burden, number of leads with T-wave inversion, and RV function. Exercise restriction is critical because catecholamine surges and mechanical stress accelerate disease progression and trigger arrhythmias.
Channelopathies
Ion Channel Diseases
The channelopathies (long QT syndrome, Brugada syndrome, catecholaminergic polymorphic VT, short QT syndrome) cause SCD in structurally normal hearts. Volume IX covers each of these conditions in depth, from the specific ion current defect to genotype-guided therapy. The substrate is electrical: altered ion channel function creates dispersion of repolarization or abnormal calcium handling that supports VT/VF initiation.
Risk stratification here is genotype-specific. In long QT syndrome, QTc duration, genotype (LQT1 vs LQT2 vs LQT3), symptom history, and sex all factor into risk. In Brugada syndrome, spontaneous type 1 ECG pattern and syncope history are the major risk markers. Each channelopathy has its own risk logic, rooted in the specific ion current that is dysfunctional.
The ICD Decision
An ICD is a defibrillator sewn into the chest. When it works as intended, it detects VT or VF and delivers therapy (antitachycardia pacing or a shock) within seconds, preventing sudden death. This is a profound intervention. It is also an intervention with real costs.
Appropriate shocks save lives. The device does exactly what it was designed to do: terminate a lethal arrhythmia before it kills the patient. Inappropriate shocks cause harm. They fire for rhythms that are not lethal: sinus tachycardia, atrial fibrillation with rapid ventricular response, lead noise, or T-wave oversensing. Each inappropriate shock delivers a painful jolt to a conscious patient. Repeated shocks cause anxiety, depression, PTSD-like symptoms, and may themselves trigger ventricular arrhythmias through catecholamine surge.
The decision to implant requires weighing the probability of a lethal arrhythmia (benefit) against the probability of device complications (harm). These complications include inappropriate shocks, lead failure, infection, and the psychological burden of living with an implanted defibrillator.
Certain situations make ICD implantation inappropriate regardless of substrate:
- Reversible cause: VT/VF that occurred during the first 48 hours of an acute MI, or from a correctable electrolyte derangement (hypokalemia, hypomagnesemia), or from drug toxicity (e.g., antiarrhythmic proarrhythmia). If the trigger is eliminated and the substrate is not independently lethal, the recurrence risk drops substantially.
- Life expectancy less than one year: The ICD prevents sudden death, but if the patient has a terminal illness with expected survival under one year, the device prolongs the dying process rather than extending meaningful life. Guidelines require at least one year of expected survival at a reasonable functional status.
- Severe non-cardiac comorbidity: End-stage renal disease, advanced malignancy, or severe dementia may shift the risk-benefit balance against implantation. Shared decision-making with the patient (or surrogate) is essential.
- NYHA Class IV heart failure not a candidate for transplant or LVAD: These patients are more likely to die of pump failure than arrhythmia. An ICD may prevent sudden death while leaving the patient to die of progressive hemodynamic collapse.
The ICD conversation is a shared decision. The electrophysiologist presents the data: the estimated annual risk of SCD, the reduction in mortality from the device, and the expected complication rate. The patient brings their values: how they weigh a possible shock against the protection from sudden death, how they feel about living with hardware in their chest, and what quality of life means to them. Neither side of this equation is purely medical.
Key Takeaways
- SCD is caused by VT degenerating into VF in the vast majority of cases. The substrate (scar, fibrosis, disarray, or channel dysfunction) allows reentry initiation; wavefront fragmentation converts organized VT into lethal VF.
- Secondary prevention patients have already demonstrated lethal substrate. ICD implantation is indicated unless the event had a fully reversible cause.
- EF ≤ 35% is the primary criterion for primary prevention ICD implantation (MADIT-II, SCD-HeFT), because low EF reflects extensive substrate. Its key limitation: most SCDs in absolute numbers occur in patients with EF above this threshold.
- Cardiac MRI with LGE directly visualizes scar and border zone tissue, providing substrate-level risk information that EF alone cannot capture. Border zone volume independently predicts VT inducibility and SCD.
- Special populations (HCM, ARVC, channelopathies) each require disease-specific risk stratification that goes beyond EF, using dedicated risk calculators and clinical features unique to each condition.
- The ICD decision balances SCD prevention against device complications. Do not implant for reversible causes, life expectancy under one year, or clinical scenarios where pump failure will precede arrhythmic death.